Hey #epitwitter, a question. At what point does the heterogeneity across epidemiological studies make the meta-analysis pointless (or at least too vague to be useful)?
-
-
For a meta analysis to be ok, don’t we need the same variables measured the same way across many studies? It seems this is difficult in epi studies. (And prob not just epi).
-
Definitely, but the suggestion is that this is impossible if we can't meta-analyse things at all. I'm not sure I entirely agree with that
- Show replies
New conversation -
-
-
This is kind of like saying unmeasured confounding on obs. studies are always an issue - true, but they're still a crucial piece of the puzzle, assuming rigorous methods & acknwlgmt of limitations.
-
See that's my perspective as well. As long as you acknowledge the issue, and recognize how it limits your study, it's usually ok I guess my question was, at what point do the limitations become greater than the rest of the study?
- Show replies
New conversation -
-
-
Meta-analytic summary estimates require a strong assumption of homogeneity across all aspects of studies (design, measurement, etc.), which is unrealistic. Meta-analysis should be used to explore sources of heterogeneity, as discussed by
@Lester_Domes. https://academic.oup.com/aje/article-abstract/140/3/290/99737 …Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Imo, I'm uncomfortable with the idea that there is ONE TRUE effect that we could get from meta-analysis. Heterogeneity is important. People respond differently to treatments. Even EXACT SAME protocol in different pop could yield diff. results. Study the het. It's not a nuisance.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.